Catálogo de publicaciones - libros

In this paper, a new hybrid weighting method, which learns region importance from the region size and the spatial location of regions in an image, is introduced to re-weight regions optimally and improve the performance of the region-based search system on the Web. Relevant images marked by an user may exhibit very different visual characteristics so that they may be scattered in several clusters in the feature space, since there exists the semantic gap between the low level feature and the high level semantics in user’s mind. Our main goal is to find semantically related clusters and their weights to narrow down this semantic gap. To do this, The hybrid region weighting method, which refines the weights of region-clusters through relevance feedback, determines the importance of regions according to the region size and spatial location information of regions in an image. Experimental results demonstrate the efficiency and the effectiveness of the proposed weighting method in comparison with the area percentage method and the region frequency weighted by inverse image frequency method, respectively.

The general problem addressed in this paper is the rapid development of web interfaces to software systems using only their command line interface. This kind of system is frequently developed in environments that greatly differ from those where web interface will be implemented. In this setting it is also important to maintain a loose coupling between the web interface and the system it controls since the latter must be able to continue its normal development independently of the former.

We propose a framework to develop web interfaces targeted to these systems whose main feature is the fact that it can be extended without requiring code programming. The hot spots of our framework are XML configuration files to define the interface data, how this data is mapped into the system’s commands, and how commands output and the interaction state is mapped into web formatting languages. With this approach the web interface is kept separated from the system it controls, it is easy to define and modify, and is able to capture enough domain knowledge to be a real advantage for the novice or sporadic user.

In this paper we present the proposed framework architecture, loosely inspired in the MVC pattern, its implementation on Java servlet containers, and its application to the AGILMAT system, a high-school mathematical problem generator developed using constrained grammars.

Using a public key certificate for authentication purpose is vulnerable to MIM attack or certificate theft. In this study, we suggest a technique to embed the user’s biometric data in all packets for a strong authentication. By authenticating each and every packet with the user’s biometric data, we can completely remove the possibility of MIM or session hijacking attack. The embedding of biometric data starts from the very first SYN packet, thereby nullifying port scanning or SYN-flooding DOS attack. The system has been implemented and shows that it can achieve an enhanced security with tolerable increase in packet length.

XML message brokers play a key role in exchanging information in ubiquitous environments. One of their core technical issues is difficulty associated with processing a set of XPath queries for multiple continuous filtering over incoming XML streams. This paper proposes a novel system designed to provide an epochal solution to this problem. The proposed system provides efficient data structures and matching algorithm in order to minimize the runtime workload of continuous filtering over XML streams. Also, the performance of proposed approach is verified through a variety of experiments, including comparisons with YFilter. The proposed approach is practically linear-scalable and stable in terms of processing a set of XPath queries in a continuous and timely fashion. Furthermore, this approach consistently outperforms YFilter, particularly under conditions of low selectivity.

With the growing of semantic web services, distributed and dynamic discovering service becomes a key requirement for service requestors. However, current service discovery mainly focuses on centralized methods, which don’t contain enough capability to support dynamic and automatic service discovery. In this paper, we propose a structured peer-to-peer semantic routing system. The peer-to-peer network and semantic matching can be used to support the completely decentralized semantic service discovery. In the presented system, with the semantic service characteristic vector, the semantic service description is distributed on structured P2P network to allow requesters routing and locating services according to semantic similarity and relationship. Finally, experimental result shows that the presented system is efficient and scalable.

In this paper, a new summarization method, which uses non-negative matrix factorization (NMF) and -means clustering, is introduced to extract meaningful sentences from multi-documents. The proposed method can improve the quality of document summaries because the inherent semantics of the documents are well reflected by using the semantic features calculated by NMF and the sentences most relevant to the given topic are extracted efficiently by using the semantic variables derived by NMF. Besides, it uses means clustering to remove noises so that it can avoid the biased inherent semantics of the documents to be reflected in summaries. We perform detail experiments with the well-known DUC test dataset. The experimental results demonstrate that the proposed method has better performance than other methods using the , the , and the .

Program slicing is a well-known technique to extract the program statements that (potentially) affect the values computed at some point of interest. In this work, we introduce a novel slicing method for XML documents. Essentially, given an XML document (which is valid w.r.t. some DTD), we produce a new XML document (a ) that contains the relevant information in the original XML document according to some criterion. Furthermore, we also output a new DTD such that the computed slice is valid w.r.t. this DTD. A prototype implementation of the XML slicer has been undertaken.

In the past few years, XML has been established as an effective means for information management, and has been widely exploited for complex data representation. Owing to an unparalleled increasing use of the XML standard, developing efficient techniques for comparing XML-based documents becomes essential in information retrieval (IR) research. Various algorithms for comparing hierarchically structured data, e.g. XML documents, have been proposed in the literature. However, to our knowledge, most of them focus exclusively on comparing documents based on structural features, overlooking the semantics involved. In this paper, we integrate IR semantic similarity assessment in an edit distance algorithm, seeking to amend similarity judgments when comparing XML-based documents. Our approach comprises of an original edit distance operation cost model, introducing semantic relatedness of XML element/attribute labels, in traditional edit distance computations. A prototype has been developed to evaluate our model’s performance. Experiments yielded notable results.

Large information spaces and complex functionality of contemporary systems together with the advent of the Semantic Web are big challenges for the design of simple yet powerful user interfaces. The complex nature of systems thus requires the use of several presentation, personalization and user modeling methods and techniques to address the growing needs and requirements of both users and systems. We present a design of an architecture for the personalized presentation layer of a web-based information system that employs a set of interconnected software components implemented as autonomous software tools for presentation, personalization and user modeling to support features such as navigation support and different views on the presented data, acquisition and evaluation of user characteristics and user adaptation and personalization. The feasibility of our design is evaluated by its realization in the labor market application domain.

Service composition technique aims to integrate several Web services to meet the increasing user demand that a single service can not satisfy. But it also brings the potential of abnormal message interactions among the composed services which lead to the service interactions problem resembling the feature interactions in the telecommunication area. Inspired by the similarities between the immune system and the service interactions detection system, this paper presents an online method to detect the service interactions problem by using ideas from the immune principles. The method overcomes the drawbacks of static methods, and conduces to an effective detection of not only known feature interactions, but also previously unknown ones.